D10-1068 |
the grammar using a supervised
|
parse selection
|
model . In the standard process
|
C02-2025 |
building statistical models of
|
parse selection
|
are possible . At the simplest
|
C02-2025 |
Table 2 shows the accuracy of
|
parse selection
|
using the models described above
|
D10-1068 |
standard process for creating a
|
parse selection
|
model is : 1 . parse the training
|
D09-1135 |
is proved to be effective for
|
parse selection
|
. Recently , Agirre et al. (
|
D10-1068 |
styles of parsing , however ,
|
parse selection
|
is based on a statistical model
|
D10-1067 |
algorithms based on quality-based
|
parse selection
|
. We first detail a basic version
|
D10-1068 |
is generally streamlined with
|
parse selection
|
models , creating the initial
|
C02-2025 |
between the parsing problem and the
|
parse selection
|
problem . The first column of
|
D10-1068 |
development of fully unsupervised
|
parse selection
|
models . The particular style
|
D09-1135 |
data to train a discriminative
|
parse selection
|
model combining syntactic features
|
D10-1068 |
completely removing this requirement of
|
parse selection
|
on explicitly treebanked data
|
D10-1068 |
effort to produce the treebank ,
|
parse selection
|
is not possible . Furthermore
|
D10-1068 |
In the standard process , the
|
parse selection
|
model is trained over a hand-disambiguated
|
C02-2025 |
they assign to them . We report
|
parse selection
|
performance as percentage of
|
C02-2025 |
tag sel . parse sel . tag and
|
parse selection
|
tasks ( accuracy ) . The results
|
C02-2025 |
compare this work with other work on
|
parse selection
|
for unificationbased grammars
|
C02-2025 |
grammar , that is , for which the
|
parse selection
|
task is nontrivial . ) We examine
|
C00-1029 |
which can be used \ -LSB- ` or
|
parse selection
|
. However , knowledge of se -
|
D10-1068 |
unsupervised . <title> Unsupervised
|
Parse Selection
|
for HPSG </title> Dridan Abstract
|